Jurnal Internasional Analisis Sem

Journal of Industrial Engineering & Management Research
Vol.2 No.4 DOI: https://doi.org/10.7777/jiemar.v2i4

http://www.jiemar.org e-ISSN : 2722-8878
Partial Least Squares Structural Squation Modeling (PLS-SEM)

Analysis for Social and Management Research : A Literature Review
Agus Purwanto1, Yuli Sudargini2
1
AGUSPATI Research Institute, Indonesia
2
SMA Negeri 1 Pati
Corresponding email : agozpor@gmail.com
ABSTRACT
Many social and management studies use quantitative methods that use partial least sound. The purpose
of this article is to explain the sequence and standard of data processing using partial least squares for
social and management research. The method of writing this article is to review articles written by Hair.
The results of the explanation of this article can help researchers to process and analyze data with partial
least squares. PLS is an acronym for Partial Least Square. Broadly speaking, PLS is a measuring tool in
statistical methods. PLS is a multivariate technique that is capable of managing various things such as
response variables to explanatory variables simultaneously. From the initial explanation, it is clear how
PLS brings convenience to those who master it.Actually, PLS is an alternative in statistics
Keywords: Partial Least Squares Structural Squation Modeling (PLS-SEM) Analysis for Social
Research : A Literature Review
Abstract
Introduction
In a study, researchers are often faced with conditions where the sample size is large enough, but has
a weak theoretical basis in the relationship between the hypothesized variables. However, it is not
uncommon to find relationships between variables that are very complex, but the data sample size is
small. Partial Least Square (PLS) is an alternative method of Structural Equation Modeling (SEM) that
can be used to overcome these problems. There are two approaches in Structural Equation Modeling
(SEM), namely covariance-based SEM (Covariance Based-SEM, CB-SEM) and SEM with variance
approach (VB-SEM) with Partial Least Squares (PLS-SEM) technique. PLS-PM has now become a
popular analytical tool with many international journals or scientific research using this method. Partial
Least Square abbreviated as PLS is a type of component-based SEM analysis with formative construct
properties. PLS was first used to process data in the field of econometrics as an alternative to SEM
techniques with a weak theoretical basis. PLS only serves as a predictor analysis tool, not a model test.
According to Hair et al.(2019) currently covariance-based structural equation modeling (CB-SEM) is
the dominant method for analyzing complex interrelationships between observed variables and latent
variables. In fact, until around 2021, far more articles published in social science and management
journals use CBSEM than partial least squares structural equation modeling (PLS-SEM). In recent years,
the number of articles published using PLS-SEM has increased significantly compared to CB-SEM (Hair
114
Electronic copy available at: https://ssrn.com/abstract=3982764

et al., 2017). PLS-SEM is now widely applied in various social science disciplines, including
organizational management (Sosik et al., 2009), marketing management (Bernarto et al.2020),
international management (Richter et al., 2015), human resource management ( Ringle et al., 2019),
management information systems (Ringle et al., 2012), operations management (Peng and Lai, 2012),
education management (Asbari et al., 2021), education management (Purwanto et al.2021) marketing
management (Hair et al., 2012), management accounting (Nitzl, 2016), strategic management (Hair et al.,
2012), hospitality management (Ali et al., 2018) and supply chain management (Kaufmann and Gaeckler,
2015 ). Several textbooks (e.g., Garson, 2016; Ramayah et al., 2016), edited volumes (eAvkiran and
Ringle, 2018; Ali et al., 2018), and special editions of scientific journals (Rasoolimanesh and Ali, 2018;
Shiau et al., 2019) describe PLS-SEM or propose methodological extensions ( Hair et al.2019).
Methode
The method of writing this article is a literature review, namely reviewing articles from several articles,
naley
1. Hair, J.F., Risher, J.J., Sarstedt, M. and Ringle, C.M. (2019), "When to use and how to report
the results of PLS-SEM", European Business Review, Vol. 31 No. 1, pp. 2-24.
Structural Equation Modeling (PLS-SEM), Sage, Thousand Oaks, CA.
2. Hair, J.F., Hult, G.T.M., Ringle, C.M., Sarstedt, M. and Thiele, K.O. (2017b), “Mirror,
Mirror on the wall: a comparative evaluation of composite-based structural equation
modeling methods”, Journal of the Academy of Marketing Science, Vol. 45 No. 5, pp. 616-
632.
3. Hair, J.F., Ringle, C.M. and Sarstedt, M. (2011), “PLS-SEM: indeed a silver bullet”, Journal
of Marketing Theory and Practice, Vol. 19 No. 2, pp. 139-151.
4. Hair, J.F., Ringle, C.M. and Sarstedt, M. (2013), “Partial least squares structural equation
modeling: rigorous applications, better results and higher acceptance”, Long Range
Planning, Vol. 46 Nos 1/2, pp. 1-12.
5. Hair, J.F., Sarstedt, M., Hopkins, L. and Kuppelwieser, V.G. (2014), “Partial least squares
structural equation modeling (PLS-SEM): an emerging tool in business research”, European
Business Review, Vol. 26 No. 2, pp. 106-121.
6. Hair, J.F., Sarstedt, M., Matthews, L. and Ringle, C.M. (2016), “Identifying and treating
unobserved heterogeneity with FIMIX-PLS: part I – method”, European Business Review,
Vol. 28 No. 1, pp. 63-76.
7. Hair, J.F., Sarstedt, M., Pieper, T.M. and Ringle, C.M. (2012), “The use of partial least
squares structural equation modeling in strategic management research: a review of past
practices and recommendations for future applications”, Long Range Planning, Vol. 45 No.
5/6, pp. 320-340.
8. Hair, J.F., Sarstedt, M. and Ringle, C.M. (2019), “Rethinking some of the rethinking of
partial least squares”, European Journal of Marketing, Forthcoming. Hair, J.F., Sarstedt, M.,
Ringle, C.M. and Gudergan, S.P. (2018), Advanced Issues in Partial Least Squares Structural
Equation Modeling (PLS-SEM), Sage, Thousand Oaks, CA.
9. Hair, J.F., Sarstedt, M., Ringle, C.M., et al. (2012b), “An assessment of the use of partial
least squares structural equation modeling in marketing research”, Journal of the Academy
of Marketing Science, Vol. 40 No. 3, pp. 414-433
115

Result and Discussion

According to Hair et al.(2019) the PLS-SEM method is of great interest to many researchers because
it allows estimating complex models with many constructs, indicator variables, and structural paths
without imposing distributional assumptions on the data. PLS-SEM is a predictive causal approach to
SEM that emphasizes prediction in estimating statistical models, the structure of which is designed to
provide causal explanations (Wold, 1982; Sarstedt et al., 2017). According to Hair et al.(2019) this
technique thus overcomes the apparent dichotomy between explanations as usual emphasized in academic
research and prediction, which is the basis for developing managerial implications (Hair et al., 2019). In
addition, there are easy-to-use software packages that generally require little technical knowledge of the
method, such as PLS-Graph (Chin, 2003) and SmartPLS (Ringle et al., 2015; Ringle et al., 2005), while
more complex for statistics Computing software environments, such as R, can also run PLS-SEM (e.g.
semPLS; Monecke and Leisch, 2012). Writers such as Richter et al. (2016), Rigdon (2016) and Sarstedt et
al. (2017) provide a more detailed argument and discussion about when to use and not to use PLS-SEM (
Hair et al .2019).
According to Hair et al.(2019) the reason the researcher chose PLS-SEM is when the analysis relates
to testing the theoretical framework of predictive perspective, when the structural model is complex and
includes many constructs, indicator or relationship models, when the aim of the research is to better
understand increasing complexity by exploring theoretical extensions of an already established theory
(exploratory research for theory development, when the path model includes one or more constructs
formatively measured, when research consists of financial ratios or similar types of data artifacts, when
research is based on secondary/archive data, which may lack comprehensive evidence on the basis of
measurement theory, when small population limits sample size, PLS-SEM also works particularly well
with large sample sizes, when distributional issues are a concern, such as a lack of normality; and when
the study requires scores of latent variables for follow-up analysis.The list above provides an overview of
points to consider when deciding whether PLS is an SEM method. appropriate for a study.
According to Hair et al.(2019) PLS-SEM offers solutions with small sample sizes when the model
consists of many constructs and a large number of items (Fornell and Bookstein, 1982; Willaby et al.,
2015; Hair et al., 2017).Technically, the PLS-SEM algorithm makes this possible by computing
measurements and structural model relationships separately, rather than simultaneously. Arstedt et al.
(2016) summarize how PLS-SEM provides a solution when methods such as CB-SEM develop
unacceptable or inconsistent results with complex and small models and sample sizes, regardless of
whether the data come from a generalized model or a composite population. According to Hair et
al.(2019) PLS-SEM can certainly be used with smaller samples but the nature of the population dictates
situations where small sample sizes are acceptable (Rigdon, 2016).Assuming that other situational
characteristics are the same, the more heterogeneous the population, the larger the sample size required to
achieve an acceptable sampling error (Cochran, 1977).
According to Hair et al.(2019) many researchers have benefited from the high level of statistics of
this method compared to CB-SEM (Reinartz et al., 2009; Hair et al., 2017). Greater statistical power
means that PLS-SEM is more likely to identify relationships as significant when they are indeed present
in the population (Sarstedt and Moi, 2019). PLS-SEM characteristics of higher statistical power are useful
for exploration research that examines underdeveloped or still developing theories.
116

The first step in evaluating PLS-SEM results involves examining the measurement model. The
relevant criteria differ for reflective and formative constructs. if the measurement model meets all the
necessary criteria, the researcher then needs to assess the structural model (Hair et al., 2017). Like most
statistical methods, PLS-SEM has a rule of thumb that serves as a guide for evaluating model results
(Chin, 2010; Götz et al., 2010; Henseler et al., 2009; Dagu, 1998; Tenenhaus et al., 2005; Roldán and
SánchezFranco , 2012; Hair et al., 2017). The reliability value for exploratory research should be at least
0.60, while the reliability for studies that depend on the established size should be 0.70 or higher.
The first step in the assessment of the reflective measurement model involves examining the loading
indicators. Loading above 0.708 is recommended, as it indicates that the construction explain more than
50 percent of the variance of the indicator, thus providing an acceptable item of reliability.The second
step is to assess the reliability of internal consistency, most often using Jöreskog's (1971) composite
reliability. A higher value generally indicates a higher level of reliability. For example, a reliability value
between 0.60 and 0.70 is considered acceptable in exploration In the study, values between 0.70 and 0.90
ranged from "satisfactory to good." Values of 0.95 and higher are problematic, as they indicate that the
item is redundant, thereby reducing construct validity (Diamantopoulos et al., 2012; Drolet and Morrison,
2001). A reliability value of 0.95 and above also indicates a possible unwanted response pattern.
According to Hair et al.(2019) researchers can use bootstrap confidence intervals to test whether
construct reliability is significantly higher than the recommended minimum (eg the lower limit of the 95
percent confidence interval of construct reliability is higher than 0.70) .Similarly, they can test whether
construct reliability is significantly lower than the maximum recommended threshold (eg the upper limit
of the 95 percent confidence interval of construct reliability is lower than 0.95). To obtain a bootstrap
confidence interval, in line with Aguirre-Urreta and Rönkkö (2018), researchers generally have to use the
percentile method. According to Hair et al.(2019) the third step of the assessment of the reflective
measurement model discusses the convergence of the validity of each construct measure. Convergent
validity is the extent to which the construct converges to explain the variance of the items. The metric
used to evaluate the constructs of convergent validity is the average extracted variance (AVE) for all
items in each construct. According to Hair et al.(2019) To calculate the AVE, we must square the loading
of each indicator on a construct and calculate the mean value. average. An acceptable AVE is 0.50 or
higher indicating that the construct explains at least 50 percent of the variance of the items.
According to Hair et al.(2019) the fourth step is to assess discriminant validity, namely the extent to
which the construction is empirically different from other constructs in the structural model. Fornell and
Larcker (1981) proposed traditional metrics and suggested that each AVE construct should be compared
with the correlation between the quadratic constructs of the same construct and all other reflectively
measured constructs in the structural model. The joint variance for all model constructs cannot be greater
than their AVE. According to Hair et al.(2019) recent research shows that this metric is not suitable for
assessing discriminant validity. For example, Henseler et al. (2015) show that the FornellLarcker criteria
do not perform well, especially when the indicators loading on the constructs are only slightly different
(eg all indicator loadings are between 0.65 and 0.85). Instead, Henseler et al. (2015) proposed a
heterotrait-monotrait ratio (HTMT) correlation (Voorhees et al., 2016). The HTMT is defined as the mean
value of item correlations across constructs relative to the mean (geometric) correlations for items
measuring the same construct. The discriminant validity problem is present when the HTMT value is
high. Henseler et al. (2015) proposed a threshold value of 0.90 for structural models with conceptually
117

very similar constructs, for example cognitive satisfaction, affective satisfaction and loyalty. In such a
setting, an HTMT value above 0.90 would indicate that discriminant validity does not exist. According to
Hair et al.(2019) But when the constructs are conceptually more different, a lower, more conservative
threshold value is suggested, such as: 0.85 (Henseler et al., 2015). In addition to these guidelines,
bootstrap can be applied to test whether the HTMT value is significantly different from 1.00 (Henseler et
al., 2015) or a lower threshold value such as 0.85 or 0.90, which should be determined based on
contextual research (Franke and Sarstedt, 2019). More specifically, the researcher can check whether the
upper bound of the 95 percent confidence interval of HTMT is lower than 0.90 or 0.85. According to Hair
et al.(2019) Variance inflation factor (VIF) is often used to evaluate the formative collinearity of
indicators. A VIF value of 5 or more indicates a critical collinearity problem among indicators
formatively measured constructs. However, collinearity problems can also occur at lower VIF values of 3
(Mason and Perreault, 1991; Becker et al., 2015). Ideally, the VIF value should be close to 3 and lower. .
According to Hair et al. (2017), indicators with insignificant weights must definitely be omitted if the
weighting is also not important. Low but significant loadings of 0.50 and below indicate that one should
consider remove indicators, unless there is strong support for their inclusion on the basis of measurement
theory.
According to Hair et al.(2019) Structural model coefficients for the relationship between constructs
are derived from estimating a series of regression equations. Before assessing the structure of the
relationship, collinearity should be checked to ensure unbiased regression of the results. According to
Hair et al.(2019) This process is similar to assessing the formative measurement model, but the latent
variable score of the predictor construct in partial regression is used to calculate the VIF value. VIF
values above 5 indicate possible collinearity problems among predictor constructs, but collinearity
problems can also occur at VIF values lower than 3-5 (Mason and Perreault, 1991; Becker et al., 2015).
Ideally, the VIF value should be close to 3 and lower. If collinearity is a problem, a frequently used option
is to create a higher order a model that can be supported by theory (Hair et al., 2017). According to Hair
et al.(2019) If collinearity is not a problem, the next step is to examine the R Square value of the
endogenous construct. According to Hair et al.(2019) R Square measures the variance, which is
described in each endogenous construct and is therefore a measure of the explanatory power of the model
(Shmueli and Kopius, 2011). R Square is also referred to as the predictive power in the sample (Rigdon,
2012). The R Square ranges from 0 to 1, with higher values indicating greater explanatory power. As a
guideline, R Square values of 0.75, 0.50 and 0.25 can be considered substantial, moderate, and weak
(Henseler et al., 2009; Hair et al., 2011). Acceptable R Square values are based on context and in some
disciplines R Square values as low as 0.10 are considered satisfactory, for example, when predicting
stock returns (Raithel et al., 2012)
According Hair et al.(2019) Researchers can also assess how deletion of certain predictor constructs
affects the f Square value of endogenous constructs. This metric is f Square effect size and is a bit
redundant with the size of the path coefficient. More precisely, the ranking order of relevance predictor
constructs in explaining the dependent construct in a structural model is often the same when compare the
path coefficient size and f Square effect size. In such a situation, f Square effect sizes should only be
reported if requested by the editor or reviewer. If the rank order of construct relevance, when describing
the dependent construct in the structural model,different when comparing path coefficient sizes and f
Square effect sizes, researchers can report f Square effect sizes to explain the presence of, for example,
118

partial or full mediation (Nitzl et al., 2016). As a rule of thumb, values higher than 0.02, 0.15 and 0.35
represent small, medium and large f effect size (Cohen, 1988).
Complex modeling often uses structural equation modeling (SEM) with various conditions that must
be met. In some modeling these conditions are sometimes difficult to fulfill. An alternative that can be
chosen while still applying complex modeling is Partial Least Square (PLS). Here I will write down some
of the reasons, a researcher chooses PLS. SEM is designed with the condition that there is strong
theoretical support, while PLS modeling can be based on (1) theory, (2) empirical research results, (3)
analogies, relationships between variables in other fields of science, (4) things normative, for example
government regulations, laws and some, (5) other rational relations. So that the theoretical basis for PLS
can be strong, weak and even exploratory.
Likewise with the measurement model, in PLS it is commonly known as the outer model, in SEM
the relationship between indicators and variables is only reflexive, while in PLS it can be reflexive or
formative. Determining indicators can be based on theory or adapting indicators that have been used by
previous researchers. Assumptions about distribution are also an important requirement in SEM. The data
in the modeling must meet the multinormal distribution, if this condition is not met then the estimate will
be shifted to resampling or bootstrapping approaches. In PLS, the assumption of a multinormal
distribution is not needed because the direct estimation uses bootstrapping technique. The sample size
needed in SEM is quite large, in many references it is recommended to have 100-200 samples. While in
PLS small samples (minimum 30-50) can be applied. Modification of the model to achieve better model
feasibility is needed in SEM, while in PLS this is not necessary. In addition to these differences, there are
several similarities between SEM and PLS, including: (1) the relationship between constructs is linear, (2)
the model can be improved by using the "trimming theory" technique, namely eliminating paths that are
not in the model, (3) a model fit calculation is required.
Reflective Measurement Model
Convergent Validity: convergent validity measures the magnitude of the correlation between
constructs and latent variables.Individual Item Reliability: individual item reliability check, can be seen
from the standardized loading factor value. The standardized loading factor describes the magnitude of
the correlation between each measurement item (indicator) and its construct. The loading factor value >
0.7 is said to be ideal, meaning that the indicator is said to be valid in measuring the construct. In
empirical research, the loading factor value > 0.5 is still acceptable. Thus, the loading factor value < 0.5
must be removed from the model (dropped). The squared value of the loading factor value is called
communalities. This value shows the percentage of constructs able to explain the variations that exist in
the indicator.
Internal Consistency or Construct Reliability: we see the internal consistency reliability of
Cronbach's Alpha and Composite Reliability (CR) scores. Composite Reliability (CR) is better in
measuring internal consistency than Cronbach's Alpha in SEM because CR does not assume the same
weight of each indicator. Cronbach's Alpha tends to lower construct reliability than Composite Reliability
(CR). The Composite Reliability (CR) interpretation is the same as Cronbach's Alpha. Limit values > 0.7
are acceptable, and values > 0.8 are very satisfactory.
Average Variance Extracted (AVE): Another measure of covergent validity is the Average Variance
Extracted (AVE) value. The AVE value describes the variance or diversity of the manifest variables that
119

can be owned by the latent construct. Thus, the greater the variance or diversity of the manifest variables
that can be contained by the latent construct, the greater the representation of the manifest variable on the
latent construct. Fornell and Larcker (1981) in Ghozali (2014:45) and Yamin and Kurniawan (2011:18)
recommend the use of AVE for a criterion in assessing convergent validity. An AVE value of at least 0.5
indicates a good measure of convergent validity. That is, the latent variable can explain the average of
more than half the variance of the indicators. The AVE value is obtained from the sum of the squares of
the loading factor divided by the error. The AVE measure can also be used to measure the reliability of
the latent variable component score and the results are more conservative than composite reliability (CR).
If all indicators are standardized, then the AVE value will be the same as the average value of block
communalities.
Discriminant Validity: the discriminant validity of the reflective model is evaluated through cross
loading, then compares the AVE value with the square of the correlation value between constructs (or
compares the square root of the AVE with the correlation between constructs). The measure of cross
loading is to compare the correlation of the indicator with its construct and constructs from other blocks.
If the correlation between the indicator and its construct is higher than the correlation with other block
constructs, this indicates that the construct predicts the size of their block better than the other blocks.
Another measure of discriminant validity is that the AVE root value must be higher than the correlation
between constructs and other constructs or the AVE value is higher than the square of the correlation
between the constructs.
Evaluation of Formative Measurement Models
There are at least five critical issues to determine the quality of the formative model, namely:Content
specification, relates to the scope of the latent construct to be measured. This means that if you want to
research, researchers must often discuss and guarantee the correct specification of the contents of the
construct.Specification indicator, must clearly identify and define the indicator. The definition of
indicators must go through clear literature and have been discussed with experts and validated with
several pre-tests.
Reliability indicator, related to the importance of the indicators that make up the construct. Two
recommendations for assessing the reliability of indicators are to look at the indicator signs in accordance
with the hypothesis and the indicator weight is at least 0.2 or significant.Collinearity indicator, states that
the indicators formed are not interconnected (very high) or there is no multicollinearity problem that can
be measured by Variance Inflated Factor (VIF). A VIF value > 10 indicates there is a problem with
multicollinearity, andExternal validity, ensures that all established indicators are included in the model.
Evaluation of the Inner Model (Structural Model)
After evaluating the construct/variable measurement model, the next step is to evaluate the structural
model or inner model.
120

The first step is to evaluate the structural model to see the significance of the relationship between the
constructs/variables. This can be seen from the path coefficient which describes the strength of the
relationship between constructs. The sign or direction in the path (path coefficient) must be in accordance
with the hypothesized theory, its significance can be seen in the t test or CR (critical ratio) obtained from
the bootstrapping process (resampling method).
The second step is to evaluate the value of R2. The interpretation of the value of R2 is the same as the
interpretation of R2 of linear regression, namely the magnitude of the variability of endogenous variables
that can be explained by exogenous variables. According to Chin (1998) in Yamin and Kurniawan
(2011:21) the R2 criteria consist of three classifications, namely: R2 values 0.67, 0.33 and 0.19 as
substantial, moderate (moderate) and weak (weak). Changes in the value of R2 can be used to see whether
the effect of the exogenous latent variable on the endogenous latent variable has a substantive effect. This
can be measured by the effect size f2. According to Cohen (1988) in Yamin and Kurniawan (2011:21) the
suggested Effect Size f2 is 0.02, 0.15 and 0.35 with exogenous latent variables having small, moderate
and large effects on the structural level.
To validate the overall structural model, Goodness of Fit (GoF) is used. The GoF index is a single
measure to validate the combined performance of the measurement model and the structural model. This
GoF value is obtained from the square root of the average communalities index multiplied by the average
value of R2 model. GoF values range from 0 to 1 with the interpretation of values: 0.1 (small GoF), 0.25
(moderate GoF), and 0.36 (large GoF).
Another test in structural measurement is Q2 predictive relevance which serves to validate the model.
This measurement is suitable if the Latin endogenous variable has a reflective measurement model. The
results of Q2 predictive relevance are said to be good if the value > which indicates the exogenous latent
variable is good (appropriate) as an explanatory variable that is able to predict the endogenous variable.
Like the analysis using CB-SEM, the analysis using PLS-SEM also uses two important stages, namely the
measurement model and the structural model. The data in the measurement model is evaluated to
determine its validity and reliability. Part of the measurement model consists of: (1). Individual loading of
each question item. (2). Internal Composite Reliability (ICR). (3). Average Variance Extracted (AVE),
and (4). Discriminant Validity.
Conclusion
PLS is an acronym for Partial Least Square. Broadly speaking, PLS is a measuring tool in statistical
methods. PLS is a multivariate technique that is capable of managing various things such as response
variables to explanatory variables simultaneously. From the initial explanation, it is clear how PLS brings
convenience to those who master it.Actually, PLS is an alternative in statistics. PLS can be used for
multiple regression analysis methods and principal component regression. Why is that? Because basically,
these two methods are immune or their scientific language is robust. Robust itself means that the model
parameters will not change much even though there are new samples taken from the total population.With
the explanation above, it is very clear that PLS provides an interesting new alternative, which can be used
for those of you who work with statistical methods because it will not bring difficult things into your
work every day.PLS does not require as many assumptions or conditions as SEM so that it provides more
meaningful convenience for you.PLS is actually used to design models only, but with the power of this
analysis, you can also use it for theory confirmation.The use of PLS can be divided into two models,
121

namely the inner model and the outer model. For the inner model, it is used for regression. While the
outer model is used to test the validity and reliability.Compared to SEM or other similar ones, PLS is
relatively simple. By using PLS, only two models are needed, namely the inner model and the outer
model. This is what makes PLS simpler and easier to learn than others.
References
Asbari, M., Novitasari, D., Pebrina, E. T., & Santoso, J. (2020). Work-Family Conflict and Employee
Performance during Covid-19 Pandemic: What is the Role of Mental Readiness to Change? JPBM
(Jurnal Pendidikan Bisnis Dan Manajemen), 6(2).
Asbari, M., Prasetya, A. B., Santoso, P. B., & Purwanto, A. (2021). From Creativity to Innovation: The
Role of Female Employees’ Psychological Capital. International Journal of Social and Management
Studies (IJOSMAS), 02(02), 66–77. https://ijosmas.org/index.php/ijosmas/article/view/18
Asbari, M., Purba, J. T., Hariandja, E. S., & Sudibjo, N. (2021). From Leadership to Innovation:
Managing Employee Creativity. Jurnal Manajemen Strategi Dan Aplikasi Bisnis, 4(1), 143–154.
Hair, J.F., Risher, J.J., Sarstedt, M. and Ringle, C.M. (2019), "When to use and how to report the results
of PLS-SEM", European Business Review, Vol. 31 No. 1, pp. 2-24. https://doi.org/10.1108/EBR-11-
2018-0203
Hair, J.F., Hult, G.T.M., Ringle, C.M. and Sarstedt, M. (2017a), A Primer on Partial Least Squares
Structural Equation Modeling (PLS-SEM), Sage, Thousand Oaks, CA.
Hair, J.F., Hult, G.T.M., Ringle, C.M., Sarstedt, M. and Thiele, K.O. (2017b),“Mirror, Mirror on the wall:
a comparative evaluation of composite-based structural equation modeling methods”, Journal of the
Academy of Marketing Science, Vol. 45 No. 5, pp. 616-632.
Hair, J.F., Ringle, C.M. and Sarstedt, M. (2011), “PLS-SEM: indeed a silver bullet”, Journal of Marketing
Theory and Practice, Vol. 19 No. 2, pp. 139-151.
Hair, J.F., Ringle, C.M. and Sarstedt, M. (2013), “Partial least squares structural equation modeling:
rigorous applications, better results and higher acceptance”, Long Range Planning, Vol. 46 Nos 1/2,
pp. 1-12.
Hair, J.F., Sarstedt, M., Hopkins, L. and Kuppelwieser, V.G. (2014), “Partial least squares structural
equation modeling (PLS-SEM): an emerging tool in business research”, European Business Review,
Vol. 26 No. 2, pp. 106-121.
Hair, J.F., Sarstedt, M., Matthews, L. and Ringle, C.M. (2016), “Identifying and treating unobserved
heterogeneity with FIMIX-PLS: part I – method”, European Business Review, Vol. 28 No. 1, pp. 63-
76.
122

Hair, J.F., Sarstedt, M., Pieper, T.M. and Ringle, C.M. (2012), “The use of partial least squares structural
equation modeling in strategic management research: a review of past practices and
recommendations for future applications”, Long Range Planning, Vol. 45 Nos 5/6, pp. 320-340.
Hair, J.F., Sarstedt, M. and Ringle, C.M. (2019), “Rethinking some of the rethinking of partial least
squares”, European Journal of Marketing, Forthcoming. Hair, J.F., Sarstedt, M., Ringle, C.M. and
Gudergan, S.P. (2018), Advanced Issues in Partial Least Squares Structural Equation Modeling
(PLS-SEM), Sage, Thousand Oaks, CA.
Hair, J.F., Sarstedt, M., Ringle, C.M., et al. (2012b), “An assessment of the use of partial least squares
structural equation modeling in marketing research”, Journal of the Academy of Marketing Science,
Vol. 40 No. 3, pp. 414-433
Purwanto, A., Purba,J.T, Bernarto,I., Sijabat,R.(2021). The Role of Transformational Leadership,
Organizational Citizenship Behaviour, Innovative Work Behaviour, Quality Work Life, Digital
Transformation and Leader Member Exchange on Universities Performance. Linguistica
Antverpiensia.2021(2).2908-2932
Purwanto, A., Purba,J.T, Bernarto,I., Sijabat,R.( 2021).Pengaruh Servant, Digital dan Green leadership
Terhadap Kinerja Industri Manufaktur Melalui Mediasi Komitmen Organisasi, Jurnal Riset Inspirasi
Manajemen dan Kewirausahaan Volume 5 No. 1 Edisi Maret 2021 Hal 1-13, DOI :
https://doi.org/10.35130/jrimk
Purwanto, A., Purba,J.T, Bernarto,I., Sijabat,R.( (2021).Peran Organizational Citizenship Behavior
(OCB), Transformational and Digital Leadership Terhadap Kinerja Melalui Mediasi Komitmen
Organisasi Pada Family Business. Jenius. 4(3). 256-
262.http://dx.doi.org/10.32493/JJSDM.v4i3.10454
Purwanto, A., Purba,J.T, Bernarto,I., Sijabat,R (2021). EFFECT OF TRANSFORMATIONAL
LEADERSHIP,JOB SATISFACTION, AND ORGANIZATIONAL COMMITMENTS ON
ORGANIZATIONAL CITIZENSHIP BEHAVIOR. Inovbiz: Jurnal Inovasi Bisnis 9 (2021) 61-69
Purwanto, A., J. T. Purba, I. Bernarto, and R. Sijabat. 2021. Effect of Management Innovation,
Transformational leadership and knowledge sharing on Market Performance of Indonesian
Consumer Goods Company. Jurnal Aplikasi Manajemen, 19(2), 424–434. Malang: Universitas
Brawijaya. http://dx.doi.org/10.21776/ub.jam.2021.019.02.18.
Peng, D.X. and Lai, F. (2012), “Using partial least squares in operations management research: a practical
guideline and summary of past research”, Journal of Operations Management, Vol. 30 No. 6, pp.
467-480. Richter, N.F., Cepeda Carri_on, G., Roldán, J.L. and Ringle, C.M. (2016), “European
management research using partial least squares structural equation modeling (PLS-SEM): editorial”,
European Management Journal, Vol. 34 No. 6, pp. 589-597.
Ringle, C.M., Sarstedt, M. and Mooi, E.A. (2010), “Response-based segmentation using finite mixture
partial least squares: theoretical foundations and an application to american customer satisfaction
index data”, Annals of Information Systems, Vol. 8, pp. 19-49.
123

Jurnal Internasional Analisis Sem

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Jurnal Internasional Analisis Sem

Uploaded by

Copyright:

Available Formats

Journal of Industrial Engineering & Management Research

Vol.2 No.4 DOI: https://doi.org/10.7777/jiemar.v2i4

Partial Least Squares Structural Squation Modeling (PLS-SEM)

Electronic copy available at: https://ssrn.com/abstract=3982764

Electronic copy available at: https://ssrn.com/abstract=3982764

Result and Discussion

Electronic copy available at: https://ssrn.com/abstract=3982764

Electronic copy available at: https://ssrn.com/abstract=3982764

Electronic copy available at: https://ssrn.com/abstract=3982764

Reflective Measurement Model

Electronic copy available at: https://ssrn.com/abstract=3982764

Evaluation of Formative Measurement Models

Evaluation of the Inner Model (Structural Model)

Electronic copy available at: https://ssrn.com/abstract=3982764

Electronic copy available at: https://ssrn.com/abstract=3982764

Electronic copy available at: https://ssrn.com/abstract=3982764

Electronic copy available at: https://ssrn.com/abstract=3982764

You might also like